1
1
Граф коммитов

12357 Коммитов

Автор SHA1 Сообщение Дата
Jeff Squyres
6b6c08ef67 Fixes trac:1588: have several BTLs disable themselves in the presence of
THREAD_MULTIPLE.  There's a new (hidden) MCA parameter to re-enable
these BTLs in the presence of THREAD_MULTIPLE:
btl_base_thread_multiple_override.  This MCA parameter should ''only''
be used by developers who are working on make their BTLs thread safe;
it should ''not'' be used by end-users!

This commit was SVN r19826.

The following Trac tickets were found above:
  Ticket 1588 --> https://svn.open-mpi.org/trac/ompi/ticket/1588
2008-10-28 18:29:57 +00:00
Jeff Squyres
96f7d971d3 Refs trac:1589: remove the OMPI_DECLSPEC because it doesn't need to be
there.

This commit was SVN r19825.

The following Trac tickets were found above:
  Ticket 1589 --> https://svn.open-mpi.org/trac/ompi/ticket/1589
2008-10-28 17:24:31 +00:00
George Bosilca
d33801c0fa Rollback the last two commits.
svn merge -r 19823:19821 https://svn.open-mpi.org/svn/ompi/trunk ./

I copied the config directory in another location, and the original
.svn directory was replaced by the Open MPI one ... As a result my
first commit, applied the changes on the Open MPI trunk instead of
the other project.

This commit was SVN r19824.
2008-10-28 17:22:29 +00:00
George Bosilca
e0bb777a51 Remove useless files.
This commit was SVN r19823.
2008-10-28 17:02:37 +00:00
George Bosilca
74041562a2 One file at the time ...
This commit was SVN r19822.
2008-10-28 16:56:53 +00:00
Kenneth Matney
4c4021cf1e Prior to these configuration files, the GNU Fortran binding (gfortran)
would not build.  These new files will force the build process to look
up the value of "logical true", instead of arbitrarily setting a value,
enabling compatibility between GNU and PGI compilers.

This commit was SVN r19821.
2008-10-28 14:51:10 +00:00
Ralph Castain
82ece176d5 Sanity check needs to allow vpid_invalid as this indicates the "none" scenario
This commit was SVN r19820.
2008-10-28 14:50:26 +00:00
Jeff Squyres
fdbb2d01aa Fix the wrapper compiler flags to get the right C++ exceptions flags
(this was missed in #1585).  Also, fix a long-standing problem that
the F90 wrapper compilers were using the F77 wrapper compiler flags.

This commit was SVN r19819.
2008-10-28 14:26:47 +00:00
Jeff Squyres
be21ea2391 Back out r19805 and replace it with something a bit better, based on a
suggestion from George during review.

Refs trac:1595.

This commit was SVN r19818.

The following SVN revision numbers were found above:
  r19805 --> open-mpi/ompi@bc710baa75

The following Trac tickets were found above:
  Ticket 1595 --> https://svn.open-mpi.org/trac/ompi/ticket/1595
2008-10-28 12:25:08 +00:00
Brad Penoff
d7b0fdfe5c small fix to compile trunk on FreeBSD 7
This commit was SVN r19817.
2008-10-28 03:44:23 +00:00
Matthias Jurenz
bd802e06c8 Undo last *faulty* checkin
This commit was SVN r19816.
2008-10-27 19:14:32 +00:00
Jeff Squyres
200a8552fe Fixes trac:1598: it's not an error if a CPC returns UNREACH.
This commit was SVN r19815.

The following Trac tickets were found above:
  Ticket 1598 --> https://svn.open-mpi.org/trac/ompi/ticket/1598
2008-10-27 18:23:56 +00:00
Ethan Mallove
2457df91b3 Add missing #include <errno.h> line (for SunStudio Solaris).
This commit was SVN r19814.
2008-10-27 17:41:33 +00:00
Tim Mattox
b081fc3462 Bring the make_dist_tarball gnu tools version checks up to date from
the v1.3 branch's r19812:

Update the make_dist_tarball script to check for the proper GNU tool
versions as posted on http://www.open-mpi.org/svn/building.php
 * m4 1.4.11
 * autoconf 2.63
 * automake 1.10.1
 * libtool 2.2.6

This commit was SVN r19813.

The following SVN revision numbers were found above:
  r19812 --> open-mpi/ompi@706923f7b1
2008-10-27 16:41:27 +00:00
Jeff Squyres
b11d13cc05 Silence a trivial compiler warnings (pgcc).
This commit was SVN r19810.
2008-10-27 14:23:02 +00:00
Jeff Squyres
c078ab6b09 Minor fix for a trivial compiler warning.
This commit was SVN r19809.
2008-10-27 14:18:49 +00:00
Matthias Jurenz
286321210a Minor bugfix in VTGen_write_COMMENT():
Call VTGEN_ALLOC_EVENT instead of VTGEN_ALLOC_DEF

This commit was SVN r19808.
2008-10-26 12:06:13 +00:00
Matthias Jurenz
1dc0348665 Fixed compiler error:
replaced 'using std::sprintf' by' using std::snprintf'

This commit was SVN r19807.
2008-10-26 12:05:56 +00:00
Matthias Jurenz
894df5072b Fixed minor portability issues under Windows
This commit was SVN r19806.
2008-10-26 12:05:27 +00:00
Jeff Squyres
bc710baa75 Remove a compiler warning from ompi_info by using 2 macros to record
whether MPI API parameters should be checked never, always, or
determined at run-time.

This commit was SVN r19805.
2008-10-25 20:54:02 +00:00
Ralph Castain
71dcf61f9b Add sanity check to ensure that specified stdin target is within range of job. Print error message and exit if not.
Modify read_write test to allow specification of rank to read stdin.

IOF now validated to work for arbitrary rank as stdin target. Not validate to work for multiple simultaneous ranks reading stdin (untested).

This commit was SVN r19804.
2008-10-25 14:38:06 +00:00
Jeff Squyres
b5123cb79f Don't destroy the event channel until after everything else has been
torn down.  Fixes trac:1582.

This commit was SVN r19800.

The following Trac tickets were found above:
  Ticket 1582 --> https://svn.open-mpi.org/trac/ompi/ticket/1582
2008-10-24 15:04:54 +00:00
Jeff Squyres
ae34fd150a It always helps to initialize a variable before you try to use it
(vs. only initializing it in some cases).

This commit was SVN r19796.
2008-10-24 14:20:07 +00:00
Jeff Squyres
d96b78fee1 If the script is there, there's no real reason to have these files in
the repo.

This commit was SVN r19795.
2008-10-24 13:42:26 +00:00
Jeff Squyres
0a741d7f81 Add scripty-foo to make the data files. Revamp the data files to be
non-uniform in content as a slightly better test.

This commit was SVN r19794.
2008-10-24 13:35:47 +00:00
Ralph Castain
c56cdac379 Finish cleanup of stdin. Set non-stdio file descriptors to non-blocking (thanks to Jeff for catching that one). Handle writes that result in "would have blocked" errno.
This commit was SVN r19793.
2008-10-24 01:42:58 +00:00
Ralph Castain
6100d88ded Cleanup the new IOF:
1. remove some stale files that were overlooked in original commit

2. add a test program and data to stress iof for stdin

3. cleanup a debug statement that caused memory corruption when reading large files

4. some minor cleanups to correctly handle xon/xoff scenarios

This commit was SVN r19792.
2008-10-23 19:11:05 +00:00
Jeff Squyres
238a6f851f Wow. And I won't name who it was who did this. You know who you are! :-)
We had a public symbol named "already_opened".  This commit changes
the name to mca_btl_base_already_opened.

I guess it's good we have the visibility stuff enabled by default!
:-)

This commit was SVN r19791.
2008-10-23 18:56:10 +00:00
George Bosilca
9260f6b157 There is no reason to ask for an ACK from the BTL here.
This commit was SVN r19789.
2008-10-22 20:13:33 +00:00
George Bosilca
7dfdf3e907 A lot of MX fixes.
1. Allow MX bonding via btl_mx_bonding MCA parameter. With this on, Open MPI
  suppose that lib MX will do the bonding, and we will only return one BTL.
  Otherwise, we return as many as devices.
2. Decrease the memory footprint, by cleaning up what we store about the
  peers and how we store it.
3. Allow multiple MX routes that share the same mapper. In this particular
  case we will link by their nic_id.
4. Allow multiple MX routes with multiple mappers. In this case we match
  the NICs based on the last 6 digits of the mapper MAC.
5. Increase the size of the eager and rendez-vous eager limits in the
  case where we are unable to register an unexpected callback with MX.
6. Increase the default max number of MX fragments.
7. Increase the max number of MX BTLs.
8. Only allow mx_if_include and mx_if_exclude if we have acess to the
  mapper.

This commit was SVN r19788.
2008-10-22 20:12:30 +00:00
Terry Dontje
a39d8d62e7 Correct the "default" message for --enable-cxx-exceptions to disabled as the
code is actually coded and correct some F90 comments to f77.

This commit was SVN r19784.
2008-10-21 20:54:08 +00:00
Jeff Squyres
ac698173b9 * Ensure to pass the C++ exceptions flags to the C and Fortran
compilers as well.   Not doing this was causing problems with
   MPI::ERRORS_THROW_EXCEPTIONS with gcc in 32 bit (but not 64 bit!). 
 * Ensure that the C and Fortran compilers actually like the C++
   exceptions flags.  If not, currently just abort.  Let's see if
   anyone complains about this -- I doubt they will because a) C++
   exception support is not enabled by default, and b) I think C++
   exceptions really only make sense within the same compiler family.

This commit was SVN r19783.
2008-10-21 20:25:20 +00:00
Tim Mattox
56c014a3a2 Resync the NEWS with the 1.2 branch.
This commit was SVN r19780.
2008-10-21 18:30:19 +00:00
George Bosilca
83474b2e1a Solve a modulus rounding error. As the modulus can be signed (in C89 it take
the sign of the divident), we have to cast the pointer to an uintptr_t in
order to be able to correctly compute how to align it on the cache line.
Rported and solved by Stephan Kramer. Thanks Stephan.

This commit was SVN r19778.
2008-10-21 17:00:39 +00:00
George Bosilca
61317cb61d Complete the r19767 commit for XGrid, i.e. allow the PLM Xgrid to build.
This commit was SVN r19777.

The following SVN revision numbers were found above:
  r19767 --> open-mpi/ompi@6e5d844c36
2008-10-21 15:37:22 +00:00
Jeff Squyres
f2a7993aa5 Refs trac:1578: Shiqing-suggested changes for valgrind configure.m4 support.
This commit was SVN r19776.

The following Trac tickets were found above:
  Ticket 1578 --> https://svn.open-mpi.org/trac/ompi/ticket/1578
2008-10-21 03:27:43 +00:00
Tim Mattox
07eda5d696 Resync the NEWS file with the 1.2 branch.
This commit was SVN r19774.
2008-10-20 15:33:39 +00:00
Ralph Castain
1cfa0c26a1 Add non-debug platform for mac
This commit was SVN r19772.
2008-10-18 13:10:05 +00:00
Ralph Castain
ebaa2c59bb Cleanup non-debug builds
This commit was SVN r19771.
2008-10-18 13:09:47 +00:00
Jeff Squyres
6d026b86b7 Fix a problem reported on the user list by Teng Lin: OPAL_PREFIX
wasn't exported in the Bourne-shell-flavor case on remote nodes.

This commit was SVN r19770.
2008-10-18 12:13:10 +00:00
Jeff Squyres
d96003fec5 Fix typo.
This commit was SVN r19769.
2008-10-18 11:52:41 +00:00
Jeff Squyres
8ea27c0ced Add a missing header file to the Makefile.am so that it can be
included in the distribution tarball.

This commit was SVN r19768.
2008-10-18 11:09:57 +00:00
Ralph Castain
6e5d844c36 Roll in the revamped IOF subsystem. Per the devel mailing list email, this is a complete rewrite of the iof framework designed to simplify the code for maintainability, and to support features we had planned to do, but were too difficult to implement in the old code. Specifically, the new code:
1. completely and cleanly separates responsibilities between the HNP, orted, and tool components.

2. removes all wireup messaging during launch and shutdown.

3. maintains flow control for stdin to avoid large-scale consumption of memory by orteds when large input files are forwarded. This is done using an xon/xoff protocol.

4. enables specification of stdin recipients on the mpirun cmd line. Allowed options include rank, "all", or "none". Default is rank 0.

5. creates a new MPI_Info key "ompi_stdin_target" that supports the above options for child jobs. Default is "none".

6. adds a new tool "orte-iof" that can connect to a running mpirun and display the output. Cmd line options allow selection of any combination of stdout, stderr, and stddiag. Default is stdout.

7. adds a new mpirun and orte-iof cmd line option "tag-output" that will tag each line of output with process name and stream ident. For example, "[1,0]<stdout>this is output"

This is not intended for the 1.3 release as it is a major change requiring considerable soak time.

This commit was SVN r19767.
2008-10-18 00:00:49 +00:00
Ralph Castain
4858c9b43c Revised platform files
This commit was SVN r19766.
2008-10-17 23:52:00 +00:00
Jeff Squyres
bccd8a6cfc Expand a few configure help string messages, mainly because Terry was
complaining that he didn't know what the defaults were ;-), but also
because it's the Right thing to do.

This commit was SVN r19765.
2008-10-17 23:05:30 +00:00
Jeff Squyres
e42139710b A typo prevented the valgrind memchecker component finding the
Valgrind header files if they weren't already in the compiler's
default header file search path.  This commit fixes that typo and adds
a little more infrastructure (via an AC_SUBST) to pass in the relevant
CPPFLAGS to the build system for the valgrind memchecker component.

This commit was SVN r19764.
2008-10-17 23:04:39 +00:00
Jeff Squyres
307d52aedc Ensure to check for internal ptmalloc2 support properly.
This commit was SVN r19760.
2008-10-17 16:27:20 +00:00
Jeff Squyres
278ad5f867 Update svn:ignore
This commit was SVN r19759.
2008-10-17 16:26:52 +00:00
Jeff Squyres
e34c93c46a Fix problem of missing ) noted by Mostyn Lewis.
This commit was SVN r19758.
2008-10-17 16:03:17 +00:00
Rich Graham
9d59c0fbd6 add comment on what ompi_convertor_need_buffers() does.
This commit was SVN r19757.
2008-10-16 15:40:21 +00:00